Multi Pose And Expression Face Images Of 1,507 Chinese Subjects

Comprising 102,476 face images from 1,507 Chinese individuals (762 males, 745 females). Each subject contributed 62 multi-pose and 6 multi-expression images, captured under varying angles, poses, and lighting conditions.

Use Cases

• Train a face recognition model on 102,476 images with identity labels for 1,507 subjects.
• Develop a pose estimation model using the 62 multi-pose images per subject.
• Build an expression classification model leveraging the 6 multi-expression images per subject.
• Analyze demographic representation across the 762 male and 745 female subjects.

Strengths

• Dataset includes 102,476 images, providing a substantial volume for model training.
• Data covers 1,507 distinct subjects, offering a moderate level of identity diversity.
• Each subject provides 68 images (62 pose + 6 expression), enabling within-subject variation analysis.

Limitations

• Dataset is a sample of a larger paid dataset, limiting its completeness.
• Geographic coverage is limited to Chinese subjects, which may reduce model generalizability to other populations.
• The specific image attributes (e.g., pose angles, expression labels) are not detailed in the available metadata.

Provenance

Source

Nexdata

Collection Method

Collected images from 1,507 individuals, with 62 multi-pose and 6 multi-expression images captured per subject.

Freshness

Last updated on 2025-04-24.

Geography

China